In this study we have for the first time investigated nemabiome in 158 samples from 61 sheep commercial farms in Sweden. The aims were to find out how nematode community structure (species composition and diversity) were influenced; i) by host age (in ewes and lambs), ii) on short-term following recent anthelmintic treatment with ABZ, IVM or LEV, and iii) on long-term in samples collected several years apart. For this we used DNA extracted from composite larval pools which then were analyzed using a PacBio NGS pipeline generating sequence data (amplicons) that were clustered into OTUs with a specifically designed analysis framework developed for eukaryotes. We identified five common OTUs representing 97% of the reads, among which H. contortus and T. circumcincta were the two dominating species. There was in general no significant differences in nemabiome post-treatment between ewes and lambs. In contrast, use of anthelmintic had a short-term dramatic effect, but not on nemabiome in post-treatment samples collected several years apart indicating there was no long-term effect.
Totally 26 OTUs were identified, of which five are well-known and globally distributed strongyle nematode species in sheep; i.e. C. ovina, H. contortus, T. circumcincta, Trichostrongylus vitrinus, Oesophagostomum venulosum. This group of big-five constituted of the majority of reads (97%). All of these had a match of ≥ 99 to 100 in query cover and identity according to BLAST searches and were therefore considered as valid species. We also recognized eight other species reported from sheep with the same confidence in most cases (Table 1). These were represented by four strongylids; Bunostomum trigonocephalum, Cooperia curticei, Nematodirus spathiger and T. axei, plus four metastrongylids (small lungworm); Cystocaulus ocreatus, Muellerius capillaris, Protostrongylus hobmaieri and P. rufescens. Among the small lungworms only C. ocreatus and M. capillaris had an identity of ≥ 99. However, combined these eight represented less than 2% of the total reads. Thus, both the epidemiological consequences and clinical impact of these minor species are probably of less importance for Swedish conditions. With the exception of T. colubriformis, which was absent in our data set, we recognized all eight major species found in UK sheep based on data generated with a similar approach [19]. However, the relative abundance of the distinctive species differed. In our study five species dominated and with H. contortus being more prevalent in Swedish compared to UK sheep. From this, it follows that there is a difference in nemabiome composition in sheep between different climatic conditions. Thus, microclimatic influences would be worthwhile to investigate especially in the light of the possible impact of climatic changes.
Besides, we identified 12 species usually not found in sheep. Among these, four species are mainly linked to cattle; Cooperia oncophora, Dictyocaulus viviparus, Ostertagia leptospicularis, and O. ostertagi [2]. However, we also identified several cyathostomins with best BLAST matches to Coronocyclus coronatus, C. labratus, Craterostomum acuticaudatum, Cylicocyclus nassatus, C. ultrajectinus, C. minutus, and Triodontophorus serratus, as well as Trichostrongylus serratus, that normally are associated with equids. To the best of our knowledge it is unknown if transmission of nematodes from equids to sheep can occur. Still, at least the matches both for Coronocyclus spp. and C. ocreatus showed high query cover and identity (≥ 99–100), indicating that these findings are valid, whereas the rest had a poor match (81–87). In contrast, opportunities for cross-infections between cattle and sheep nematodes have been studied through experimental cross-infections [29]. It was shown that both Cooperia spp. and Ostertagia spp. can infect lambs, however there is varying degrees of host specificity with C. oncophora being more adapted to cattle than sheep, whereas O. ostertagi is more strongly adapted to cattle. However, as stated before by Herlich (1971), it cannot be precluded that small numbers of O. ostertagi may attain sexual maturity in sheep under natural grazing conditions, which is in line with our findings. The presence of Dictyocaulus viviparus is more obscure. However, the query cover and identity were only 97% indicating that this OTU is likely another species; probably the lungworm of sheep, D. filaria. These species are similar but genetically distinct and host specific [31]. We also identified an OTU with best match to Strongyloides fuelleborni in two samples. This species is usually found in primates and occasionally also in humans [32], whereas S. papillosus infects ruminants [2]. This OTU had a query cover 100%, but only an identity match of 90% (Table 1). Like with D. viviparus this illustrates a difficulty with interpretation of data generated through nemabiome analysis. Since, cross infections of worms in this genus between unrelated hosts is uncommon, we probably dealt with S. papillosus [33]. This species has received less attention than those in humans from a sequencing point of view, and it is therefore not available in common databases.
Interestingly, we recognized only one species,Trichostrongylus retortaeformis (OTU 100/98), primarily associated with wildlife, which usually is found in lagomorphs [2]. As outlined above there are also several other species listed herein that are primarily parasites of sheep, that are generalist known to infect a wide range of wildlife ungulates in Europe [7]. Not at least roe-deer in Spain [34], Italy [35], Turkey [36], and France [37], is known to be infected with several of the species identified herein. Among these, particularly H. contortus is the most pathogenic nematode of sheep in Sweden [3] and is therefore of major interest. Although, roe-deer is abundant in Sweden, knowledge about its nemabiome and its role as a reservoir of livestock parasites is presently unknown. Thus, this needs to be studied in the future, especially with focus on H. contortus, not at least in face of climatic change and its propensity to develop resistance to anthelmintic drugs. Since it is known that there are both polymorphisms and cryptic species around, particularly among members of the members in superfamily Trichostrongyloidea that hybridize [38], we are convinced that nemabiome analysis is well suited for this.
Clearly, the culturing conditions for nematode eggs can have an impact the nemabiome composition. For example, it has been shown that fewer eggs of T. circumcincta develop to the third stage than for T. colubriformis [39]. Besides from the input material, the number of recognized OTUs are also influenced; i) by primer design and PCR conditions, ii) the choice of NGS platform and bioinformatics pipeline for OTU clustering and recognition, and not at least iii) available sequence information in the sequence databases. In this study we took advantage of the commonly used primers NC1 and NC2 and compared the obtained sequence with information in NCBI. In line with previous studies, these universal primers spanning the first (ITS1) and second (ITS2) internal transcribed spacers in the nuclear ribosomal DNA array not only amplify a wide range of livestock genera of nematodes of veterinary interest (i.e. Bunostomum, Chabertia, Cooperia, Dictyocaulus, Haemonchus, Nematodirus, Oesophagostomum, Ostertagia, Protostrongylus, Teladorsagia, and Trichostrongylus), but also provide suitable genetic markers for species delimitation (for a review see Gasser et al., 2008). In our study, all these genera were found as well as six additional genera (Table 1.). Still, it is unclear where to define the cut-off for discriminating between nematode species, as genetic isolation is generally used to define species boundaries rather than DNA differences [40]. Nevertheless, in agreement with previously studies on nemabiome in livestock and horses [16–20], a cut-off identity threshold of ≥ 99 seems reasonable. However, as shown in our study, the query cover also needs to be considered. It was beyond the scope of the present study to investigate factors that may introduce sequencing bias. Still, when targeting the ITS region it has been shown that PacBio reflect the composition better than Illumina MiSeq for metabarcoding of fungal communities due to sequence length variation [41]. In this study we have shown that ITS2 for the different OTUs varied between 266 and 512 bp (Table 1). However, to what extent nemabiome composition is affected by sequence length variation needs to be explored.
Although no major effects in relation to host age or long-term effects were observed when we compared the samples collected 2007 and 2014–2016 (Figs. 3 and 4), it is clear that use of anthelmintic drastically influenced the nemabiome composition (Fig. 4). In this context it is important to consider that the use of the tested drugs has changed during the last decades. In Sweden, the drug of first choice whenever H. contortus is present, has from the 1960’s until recently changed from ABZ to IVM [42]. However, with emerging evidence for double resistant H. contortus strains, this practice has changed. Today, LEV is increasingly used, but currently only on farms that neither respond adequately to IVM nor ABZ. Interestingly our data show that the identified nematode species responded differently to these drugs. Of particular practical interest is that both IVM and ABZ were unable to control H. contortus on several farms, whereas LEV still has a reasonable efficacy. However, despite there was always zero-egg counts post-treatment with LEV it is evident that in particular H. contortus survived at a low level. Likewise, T. circumcincta, which is the second most important nematode in Swedish sheep, survived treatment with either ivermectin or ABZ, but then to a lesser extent than H. contortus, whereas not at all after LEV treatment. This is partly in contrasts to the situation reported several years ago, when resistance to LEV was widely prevalent among several trichostrongylid nematodes of sheep [43]. At large we believe the observed patterns reflect the current anthelmintic resistance situation in GIN of sheep in Sweden. Of major practical concern is whether LEV resistant H. contortus will appear with increasing use of this drug. In countries with more intense sheep production (i.e. New Zealand), LEV is considered as an old drug. However, in the past when this drug was more widely used in sheep, high levels of resistance was present, like for most other commonly used anthelmintics [44]. In the light of this, our observation showing low levels of surviving H. contortus is of great concern.
In conclusion, in our study, the nemabiome approach proved to be a powerful method of studying nematode community diversity in sheep and how it is influenced by factors such as host age and the most recent use of anthelmintic treatment. This approach has a high level of sensitivity and specificity, indicating that previous estimates of diversity of the sheep nemabiome may have been underestimated. More importantly it is also clear that the effects of recent anthelmintic treatment can be investigated with high precision and therefore provides more detailed information than generated by the more traditional parasitological methods such as egg counts and morphological identification of larvae.